John Snow: The Pump and the Register

In 1853 John Snow published a paper about giving chloroform during childbirth. He distinguished relief from total unconsciousness and treated the condition of the individual patient as part of the decision. The paper belongs to a practice in which a physician had to attend to changes as they happened. It also belongs to its own century. Its importance here lies in the distinctions Snow was making, and in the fact that the physician later remembered for a water pump had another demanding occupation.

Snow lived from 1813 to 1858. The surviving publications let us follow two strands of his working life without pretending to know what he thought on any particular morning. In the anesthesia paper a single patient remains in view while the physician considers the consequences of an intervention. In the cholera writings the field is a city, and the people who might have shared an exposure need to be identified after they have gone home or died. The scale changes what a physician can see. At a bedside a change may be apparent as it occurs. Across several streets the same kind of attention requires other people to collect and keep records.

Snow's water hypothesis preceded the Broad Street outbreak. He had published an earlier edition of his work on cholera in 1849, and the enlarged edition appeared in 1855. His argument connected material passed by people with cholera to what other people might ingest. Water provided one way for that material to reach many people. The proposal gave him questions to ask, but it could also lead him to favor some answers. The value of his subsequent inquiry depends in part on how it dealt with evidence that did not fit the easiest version of his idea.

Henry Whitehead, a local curate, later recalled being asked to visit a house in the Berwick Street district on the morning of September 1, 1854. Four people there had fallen ill overnight. As he went through the neighborhood, he encountered more sick people. Other parish workers reported the same experience. Whitehead's account begins with visits made to people in trouble, before those visits had become information in a report.

His later reconstruction distinguishes the day someone fell ill from the day they died. The difference matters. A rise in recorded deaths can follow an earlier rise in the number of people falling ill, and the two curves need not peak together. Whitehead remembered the change in severity and frequency as becoming apparent after several days. Once the records were assembled, he saw a decline in fatal attacks beginning earlier than the people attending patients had perceived. The person on the street and the person looking at the completed record were the same man at different stages of the inquiry.

Snow approached the outbreak with the question of shared water already in mind. The public pump in Broad Street was an obvious place to investigate because so many deaths were reported nearby. An address could take him to a door, but the address alone could not say what had been carried through it. Water could be fetched by one person for a household, served at work or taken somewhere outside the immediate neighborhood. A person registered at one address might have spent the working day at another. The inquiry needed both the reach of the register and the details that people could supply face to face.

In the map accompanying Snow's 1855 book, short dark marks accumulate along the streets around Broad Street. The marks refer to deaths at recorded locations. They make the concentration legible in a way a column of street names does not. They also create the temptation to read the picture as if it contained the whole inquiry.

Try reading the map with one hand covering the pump. The concentration of marks is still there. Now uncover the pump and ask which people in those streets actually drew their water from it. The map cannot answer that second question on its own. Even the distance between a house and a pump can mean several things. A straight line is useful on paper, while a person carrying a pail must follow a route. A nearer pump may also be less convenient on the way back from work. At the scale of a household, the relevant geography consists of habits and errands as well as buildings.

The parish inquiry published its own report with a map associated with Snow's contribution. The 1855 book and the parish report belong to related efforts, but it is worth keeping their settings distinct. One is a physician's sustained case for a mechanism of transmission. The other is a local investigation presented to the vestry, with contributions from people who approached the outbreak from different roles. The marks are easiest to interpret when read alongside those records. They locate what was recorded and direct attention to what still had to be asked.

A workhouse in Poland Street stood close to streets with heavy mortality. Snow reported five deaths among its 535 inmates, distinguishing them from people admitted after falling ill. The institution had its own water supply. It offered a comparison within the affected area, where broad explanations about the neighborhood had to account for sharply different experiences at nearby addresses.

Snow also contacted the proprietor of a brewery near the pump after noticing that the register did not show deaths among its brewery workers. The proprietor told him about the workers' drinking arrangements and the brewery's own supply. Farther along Broad Street, Snow described a percussion-cap manufactory where tubs were supplied from the street pump. Eighteen workers, he reported, died at their homes. Looking only at their home addresses would scatter across the map a group linked by a possible exposure at work.

These accounts differ in what they can establish. A proprietor's statement concerns a workplace and its habits. A workhouse count concerns a defined group of residents, with a distinction between existing inmates and new admissions. A worker who died at home required the investigator to connect two locations. Snow used each example to make the proposed connection between water and illness more specific. The workhouse and brewery gave him places to ask why the surrounding pattern weakened. The manufactory gave him a reason to look for a connection that the death addresses concealed.

Snow took his concern about the pump to the local authorities in early September. The handle was removed. It was a concrete intervention. People could no longer draw water by working that handle, and a suspected route of exposure had been interrupted at that point.

The timing of the decline is a separate matter. Whitehead's later reconstruction places the decline in fatal attacks before the removal. Snow himself acknowledged how far the outbreak had already subsided. A decision to interrupt a suspected exposure can make sense on the evidence then available. Whether that decision caused the observed decline is a different question, and the chronology cannot assign the whole decline to the handle.

Whitehead continued collecting local evidence. The parish inquiry brought together street-level testimony and an examination of the well and nearby drainage. The later investigation could examine how contamination might have reached the water, a question left open even when Snow had persuaded officials to act. Whitehead's own narrative also gives a reader access to the early experience of the outbreak and the way he revised his impression of its timing. The parish record is valuable partly because its investigators could return to questions that an urgent decision had left unanswered.

While Broad Street supplied an unusually concentrated cluster of deaths, South London offered a different kind of comparison. Two companies served households in overlapping districts. Lambeth had moved its intake upstream, while Southwark and Vauxhall continued to draw from a more polluted reach of the Thames during the period Snow studied. In some places customers of the two companies lived close to one another. The useful question was which supply reached a household, even where the address gave no immediate answer.

Snow's published study describes identifying suppliers and extending the inquiry through district registrars after William Farr became interested in the findings. Even with that help, some addresses were hard to classify. Death could occur in a workhouse after illness began elsewhere. Snow explicitly discussed that source of error. His later paper reported about 107 cholera deaths per 100,000 estimated inhabitants among the customers of Southwark and Vauxhall, compared with about eight among Lambeth customers. Those are figures for the period and population estimates in that paper. Earlier subsets in his investigations cover different intervals and should be read with their own denominators.

A comparison like this takes labor that the final pair of numbers can conceal. Each fatal case has to be attached to the place where the relevant exposure may have occurred. The number of people served by each company has to be estimated. When one classification changes, the effect can differ between the two groups. The proximity of many customers makes the comparison more informative, because broad features of the district are shared. The assignment of suppliers still came from the city's history and plumbing, rather than a random allocation for the purpose of an experiment.

Snow's work on anesthesia remained visible in print after his death in 1858. A contemporary review of his posthumously published book on chloroform describes its relation to earlier work and articles. Read beside his cholera publications, it restores a working life to a figure often identified with a single act. Snow wrote for people who had to make decisions while some consequences were still uncertain. For the anesthetist the immediate evidence came from the patient. For the investigator tracing cholera it arrived in forms as different as an address, a visit and the name of a water company.

The register was useful because it could outlast the immediate scene. It could be revisited after memory and the first urgency had passed. Its limits also became apparent when someone moved between home and work or entered an institution after falling ill. Whitehead's early visits, the brewery conversation, the workhouse count and the registrars' returns supplied pieces that no single record contained. Snow's contribution was to bring these pieces into a comparison that other people could inspect.

Return, finally, to the row of marks around the pump. We can now read it in two directions. One leads from the cluster toward a possible common source and an action local officials could take. The other leads outward to the homes, workplaces and water supplies that give the marks their meaning. The handle belongs in the first account. The register belongs in both. It helped make the losses visible, and it gave the people who came back to the question a place to begin.

John Snow, On the Mode of Communication of Cholera, second edition, 1855. Read the Golden Square inquiry for the workhouse, brewery, manufactory and removal chronology. The digitized Wellcome copy is at https://wellcomecollection.org/works/uqa27qrt and a readable transcription is at https://www.academia.dk/MedHist/Biblioteket/Print/snow_1855.html.

The Cholera Inquiry Committee, Report on the Cholera Outbreak in the Parish of St. James, Westminster, 1855. The Wellcome Collection holds the report and separately catalogues the map associated with Snow's contribution. See https://wellcomecollection.org/works/z8xczc2r and https://wellcomecollection.org/works/uxgfjt62. Henry Whitehead's later firsthand account, The Broad Street Pump, is available at https://wellcomecollection.org/works/pv2k6z8x/items. Its opening pages distinguish fatal attacks from deaths and describe his initial visits.

John Snow, Cholera and the Water Supply in the South Districts of London in 1854. The original study is transcribed at https://www.gutenberg.org/cache/epub/66507/pg66507-images.html. Its reported mortality rates use estimated populations served by each company and should be read together with its discussion of classification errors. For the other strand of Snow's work see his 1853 paper, On the Administration of Chloroform During Parturition, preserved at https://www.woodlibrarymuseum.org/rare-book/snow-j-on-the-administration-of-chloroform-during-parturition-1853/, and the contemporary review Snow on Chloroform at https://pmc.ncbi.nlm.nih.gov/articles/PMC5184002/. The book is a selective historical biography. It reconstructs arguments from the published record without invented dialogue or clinical instruction.